259 



point where the premaxillary curves rapidly backward, the most 

 lateral tooth of the bone lies immediately mesial to the pore, reckon- 

 ing along the anterior edge of the bone. This fourth infraorbital 

 pore thus lies betw^een the toothed and untoothed parts of the bone, 

 and it certainly marks the lateral limit of that part of the bone that 

 is homologous with the premaxillary of Amia, and also of that of 

 many, if not all Teleosts. 



The ventral layers of this anterior, toothed part of the premaxil- 

 lary bone of Polyterus, excepting only at its mesial end, are produced 

 backward and mesially, and form a thin, flat process which rests 

 directly upon, and is immoveably attached to, the ventral surface of 

 the cartilage that forms the floor of the nasal chamber. This process, 

 which may be called the palatine process of the bone, is covered 

 orally by the vomer of Traquair's descriptions, that bone being so 

 loosely attached to it, by loose connective tissues, that it is freely 

 moveable upon it. The process of the premaxillary can be roughly 

 said to be triangular in shape, with a deep notch cut out of its mesial 

 edge where that edge joins the tooth-bearing part of the bone. This 

 gives to the plate-like process the appearance shown in Fig. 1, the 

 processes of opposite sides of the head enclosing a small oval or 

 lozenge-shaped space, in which the cartilage of the basis cranii is ex- 

 posed on the roof of the mouth cavity. Traquair says that the 



ECP 



Fi?. 2. Fig, 3. 



•"ig- 2. Ventral view of the right maxillary, ectopterygoid and breathing- valve 

 bones. 



Fig. 3. Dorsal view of the same bones. 



17* 



